CN102692498B

CN102692498B - Automatic code reading device and automatic code reading method of sulfanilamide type medicine residue detecting reagent strip

Info

Publication number: CN102692498B
Application number: CN201210189660.XA
Authority: CN
Inventors: 王巧华; 丁幼春; 刘亚丽; 芦茜; 许堃瑞
Original assignee: Huazhong Agricultural University
Current assignee: Huazhong Agricultural University
Priority date: 2012-06-11
Filing date: 2012-06-11
Publication date: 2014-11-12
Anticipated expiration: 2032-06-11
Also published as: CN102692498A

Abstract

The invention discloses an automatic code reading device and an automatic code reading method of a sulfanilamide type medicine residue detecting reagent strip and relates to the field of sulfanilamide type medicine residue detection. The method comprises the following steps that 1, materials are prepared; 2, the reagent strip is placed in an illumination box of the device, and images are collected through a charge coupled device (CCD) camera and an image collecting card; and 3, data is recorded through a computer, and images are processed according to the following steps that A, collected colored images are converted into gray level images; B, the gray level images are enhanced; C, the characteristic parameters of the gray level images are extracted; and D, the characteristic parameters are substituted into a back propagation (BP) neural network recognition model, and the sulfanilamide type medicine residue concentration is judged through peak height detection. The machine vision technology is adopted, and the automatic code reading can be completed, the standards are unified, the judgment is accurate, the computer is used for processing images, the efficiency is high, a machine is used for replacing the human eyes, the requirements of the modern automatic detection industry can be met, and good application and popularization prospects are realized. The sulfanilamide type medicine concentration can be fast and accurately detected.

Description

Sulfa drug residue detects the automatic code-reading apparatus of reagent strip and method thereof

Technical field

The present invention relates to sulfa drug residue detection field, relate in particular to a kind of sulfa drug residue and detect the automatic code-reading apparatus of reagent strip and method thereof; Specifically, the present invention is a kind of medicine automatic identification technology, the reagent strip that detection material has been added in utilization presents different colors from sulfa drugs reaction, detect rapidly and accurately having or not of sulfa drugs and concentration by Vision Builder for Automated Inspection again, for the food that stops sulfa drugs content overproof comes into the market to provide one to have detection means fast and effectively.

Background technology

Sulfa drugs (sulfonamides is called for short SAs) is the artificial synthetic antibacterials of a class, in animal husbandry is produced, is widely used.But because the unreasonable use of SAs causes it residual in animal derived food, and enter human body by food chain, can cause human body bad reaction and there is potential carcinogenicity, brought potential threat to health; As increase the drug resistance of bacterium, make to use the result for the treatment of of medicine of the same race reduce and body is produced to toxic action under morbid state.Sulfadimidine (SM2) etc. can bring out rodent thyroid gland hyperplasia and thyroid follicular cells gland cancer.Residue limits in Food and Drug Administration's (Food and Drug Administration is called for short FDA) regulation animal derived food is 100ng/mL, must not detect sulfamethoxazole (SMZ).China specifies that the maximum residue limit(MRL) of sulfa drugs in animal derived food is 100ng/mL, and SM2 is 25ng/mL, and various sulfa drugs total amount must not exceed 100ng/mL.

At present, automatically also belong to the exploratory stage aspect identification detecting reagent strip based on machine vision medicament residue both at home and abroad, be badly in need of development.

Summary of the invention

The object of the invention is to adopt new technical scheme, provide a kind of sulfa drug residue to detect the automatic code-reading apparatus of reagent strip and method thereof.

The object of the present invention is achieved like this:

The present invention utilizes sulfamido to produce different colors from the detection material reaction on reagent strip, with the image of ccd video camera picked-up reagent strip, be sent to computing machine by image pick-up card, computing machine carries out a series of image processing, then sets up automatic identification model according to the correlation parameter of statistics and setting.

The present invention mainly completes following several work:

1, the image that gathers sulfamido colloidal gold immuno-chromatography test paper strip, mainly comprises the preparation of sulfamido colloidal gold immuno-chromatography test paper strip;

2, find suitable algorithm, carry out image processing, extract characteristic parameter;

3, design neural network, sets up automatic identification model.

Specifically, technical scheme of the present invention is as follows:

One, sulfa drug residue detects the automatic code-reading apparatus of reagent strip (abbreviation device)

This device is made up of lighting box, objective table, CCD camera, annular daylight lamp, image pick-up card and computing machine;

Bottom at lighting box is provided with objective table, in the middle of the top of lighting box, be built-in with annular daylight lamp, be provided with CCD camera at the top of lighting box, the camera of CCD camera is just being opposite to the test agent bar on objective table through the middle circular hole (about 40mm) at lighting box top;

CCD camera, capture card and computing machine are connected successively; Image pick-up card is arranged in the PCI slot of computing machine.

Two, sulfa drug residue detects the automatic reading code method of reagent strip (abbreviation method)

This method comprises the following steps:

1. material is prepared

Prepare 1 sulfa drug residue and detect reagent strip (abbreviation reagent strip)

It is sulfamido colloidal gold immuno-chromatography test paper strip that described sulfa drug residue detects reagent strip.

It is by sample pad, pad, nitrocellulose filter, adsorptive pads and PVC(polyvinyl chloride, Polyvinylchloride) backing formation.PVC backing one end adheres to sample pad, pad successively, the middle nitrocellulose filter that adheres to, and the other end adheres to adsorptive pads, it is characterized in that being coated with on pad anti-sulfa drugs monoclonal antibody specific-colloid gold label thing; On nitrocellulose filter, coated N-sulfanilamide (SN)-PABA-carrier protein couplet thing, as detection line, has been coated with rabbit anti-mouse igg as nature controlling line.

In the time that on reagent strip p-wire, antibody or antigen mate with sulfa drugs, p-wire colour developing is shallow; In the time that antibody or antigen do not mate, p-wire colour developing is dark, therefore the concentration information of reagent strip colour developing depth reaction sulfa drugs solution.

Refer to: the immunity colloidal gold test paper strip that detects sulfa drug residue

The applying date: 2006-07-17

The patent No.: ZL 200610019642.1

Patentee: Hua Zhong Agriculture University

Inventor: Bi Dingren etc.

Granted publication number: CN 100396774C

Granted publication day: 2008-06-25.

2. the lighting box of reagent strip being put into this device is by CCD camera and image pick-up card collection image;

3. by the good data of computer recording and image is processed:

A, the coloured image of collection is converted to gray level image

Conventionally preserve with the Nonlinear Scale of 8 of each sampled pixel for the gray level image showing, can have like this 256 grades of gray scales, this precision just can have been avoided visible Distortion of Striping, and is highly susceptible to programming;

B, to gather figure image intensifying

By the conversion of gradation of image grade is emphasized the feature that is concerned to suppress not concerned feature to reach, picture quality is improved, quantity of information is enriched;

The characteristic parameter of C, extraction image

Reagent strip gray level image is asked to row pixel and (finding: the peak height figure of variable concentrations has certain rule);

D, substitution BP neural network recognization model, judge the concentration of sulfa drugs by detection peak height.

Described BP neural network recognization model is the model that utilizes a kind of disclosed Multi-layered Feedforward Networks algorithm of training by Back Propagation Algorithm.

Principle of work of the present invention:

After sulfamido colloidal gold immunochromatographimethod reagent strip is reacted with determinand, put into lighting box, turn on annular daylight lamp, for sulfamido colloidal gold immunochromatographimethod reagent strip provides light source, obtain the image of reagent strip with CCD camera; Computer is built-in with image pick-up card and image processing software, can set up recognition mode and come the characteristics of image of identification agent bar, judges the concentration of sulfa drugs by characteristics of image.

The present invention has the following advantages and good effect:

1, adopt machine vision technique, can complete automatic reading code;

2, standard is unified;

3, judgement accurately;

4, utilize computing machine to process image, efficiency is high;

5, replace human eye with machine, can adapt to modernization and automatically detect industry, have a good promotion prospects.

The present invention can realize quickly and accurately and detecting for the concentration of sulfa drugs; Known through testing, detect accuracy and can reach 96%.

Brief description of the drawings

Fig. 1 is the structural representation of this device, in figure:

00-test agent bar;

10-lighting box; 20-objective table; 40-CCD camera;

30-annular daylight lamp; 50-image pick-up card; 60-computing machine.

Fig. 2 is the block diagram of this method.

Fig. 3 is the process flow diagram that extracts characteristic parameter (HIS).

Fig. 4 is the process flow diagram that extracts characteristic parameter peak height.

Embodiment

Describe in detail below in conjunction with drawings and Examples:

One, device

1, overall

As Fig. 1, this device is made up of lighting box 10, objective table 20, CCD camera 40, annular daylight lamp 30, image pick-up card 50 and computing machine 60;

Be provided with objective table 20 in the bottom of lighting box 10, in the middle of the top of lighting box 10, be built-in with annular daylight lamp 30, be provided with CCD camera 40 at the top of lighting box 10, the camera of CCD camera 40 is just being opposite to the test agent bar 00 on objective table 20 through the middle circular hole (about 40mm) at lighting box 10 tops;

CCD camera 40, capture card 50 and computing machine 60 are connected successively; Image pick-up card 50 is arranged in the PCI slot of computing machine 60.

2, functional block

1) annular daylight lamp 30

Annular daylight lamp 30 is a kind of general outsourcing pieces, as selects Philips TLSC type annular daylight lamp;

Its function is to provide light source for sulfa drugs detects reagent strip 00.

2) CCD camera 40

CCD camera 12 is a kind of general outsourcing pieces, as selects UI-2210RE-C-HG industrial camera;

Its function is the view data that gathers sulfa drugs detection reagent strip 00.

3) computing machine 60

Computing machine 60 is a kind of general outsourcing pieces, as selects association's main frame (CPU frequency 384MHZ, internal memory 128M), is built-in with image pick-up card 50 and image processing software.

The workflow of described image processing software is:

(1), as Fig. 3, extract the flow process of characteristic parameter (HIS):

1. input test agent bar image 301;

2. image pre-service: the enhancing of reagent strip image with cut apart 302;

3. input the sequence number 303 of test agent bar;

4. intercept the middle part 304 in reagent strip p-wire region;

5. extract characteristic parameter H, S, I 305;

H: tone S: saturation degree I: brightness

6. save data 306;

7. judge whether to continue 307, be to turn to jump to step 2. 302, otherwise enter next step;

8. termination routine 308.

(2), as Fig. 4, extract the flow process of characteristic parameter peak height:

1. input test agent bar image 401;

2. image pre-service: the enhancing of reagent strip image, denoising with cut apart 402;

3. input the sequence number 403 of test agent bar;

4. find reagent strip p-wire region (deducting 50 by the capable pixel of gray level image and peaked index value) 404;

5. extract characteristic parameter crest height value 405;

6. save data 406;

7. judge whether to continue 407, be to turn to jump to step 2. 402, otherwise enter next step;

8. termination routine 408.

3, the reagent strip BP neural network design of identification automatically

1) definition of sample data.

Medicament residue reagent strip is divided into five grades according to its concentration: 0 concentration (0ng/mL), 10 concentration (10ng/mL), 20 concentration (20ng/mL), 40 concentration (40ng/mL), 80 and 100 concentration (80ng/mL and 100ng/mL)

2) importing of sample data

Owing to extracting when characteristic parameter, characteristic parameter is saved in excel form, so need first the data importing in excel to MATLAB.

3) write network program

Write matlab program, its workflow is:

1. input tutor's signal and training mode;

2. setting network level and node;

3. training network;

4. train and successfully preserve network data, otherwise re-start 2..

4) network verification

The pictorial information substitution BP neural network of previous processed is carried out to computing, obtain a result and contrast with actual drug concentration, the accuracy of inspection BP neural network.

4, testing result

Judicious total reagent strip number is 58, and the 0 concentration reagent strip automatically accuracy of identification is 100%; The 10 concentration reagent strips automatically accuracy of identification are 80%; The 20 concentration reagent strips automatically accuracy of identification are 100%; The 40 concentration reagent strips automatically accuracy of identification are 100%; 80 and the 100 concentration reagent strips automatically accuracy of identification are 100%.Therefore medicament residue reagent strip concentration automatically total accuracy of identification is 96% left and right.

Claims

1. sulfa drug residue detects the automatic reading code method of reagent strip,

Device is made up of lighting box (10), objective table (20), CCD camera (40), annular daylight lamp (30), image pick-up card (50) and computing machine (60);

Be provided with objective table (20) in the bottom of lighting box (10), in the middle of the top of lighting box (10), be built-in with annular daylight lamp (30), be provided with CCD camera (40) at the top of lighting box (10), the camera of CCD camera (40) is through the middle circular hole at lighting box (10) top, be just opposite to the test agent bar (00) on objective table (20);

CCD camera (40), capture card (50) and computing machine (60) are connected successively; Image pick-up card (50) is arranged in the PCI slot of computing machine (60);

It is characterized in that comprising the following steps:

1. material is prepared (201)

Prepare 1 sulfa drug residue and detect reagent strip;

2. the lighting box of reagent strip being put into this device is by CCD camera and image pick-up card collection image (202);

3. by the good data of computer recording and image is processed to (203):

A, the coloured image of collection is converted to gray level image (204);

B, to gray level image strengthen (205);

The characteristic parameter (206) of C, extraction gray level image

To reagent strip gray level image ask row pixel and, obtain the characteristic curve of gray level image, peak concentration height is wherein characteristic parameter;

D, substitution BP neural network recognization model, judge the concentration (207) of sulfa drugs by characteristic parameter H, S, I and peak height;

Described extraction characteristic parameter H, the flow process of S, I:

A, input test agent bar image (301);

B, image pre-service: the enhancing of reagent strip image with cut apart (302);

The sequence number (303) of c, input test agent bar;

A part (304) in the middle of d, intercepting reagent strip p-wire region;

E, extraction characteristic parameter H, S, I (305);

H: tone; S: saturation degree; I: brightness;

F, save data (306);

G, judge whether continue (307), be to turn to jump to step b (302), otherwise enter next step;

H, termination routine (308);

The flow process of described extraction characteristic parameter peak height:

I, input test agent bar image (401);

II, image pre-service: the enhancing of reagent strip image, denoising with cut apart (402);

The sequence number (403) of III, input test agent bar;

IV, find reagent strip p-wire region (404);

V, extraction characteristic parameter crest height value (405);

VI, save data (406);

VII, judge whether continue (407), be to turn to jump to step II (402), otherwise enter next step;

VIII, termination routine (408).